Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-03-15 · source-backed
Standard retrieval benchmarks actively mispredict agent memory performance. Larger 10B embedding models often lose to 300M models on memory tasks. The first benchmark exposing this fundamental evaluation gap. arXiv 2603.12572
Each link below shares sources, entities, or timing with this story.
Standard evaluation of frozen-embedding style classification uses random splits where works by the same artist appear on both sides (arXiv 2608.14435). Under an artist-disjoint protocol on 320 paintings across four twentieth-century movements, 5-NN accuracy falls ten points, a...
OpenAI published "Path to Astra: critical capabilities and frontier safeguards" on September 1, declaring Astra the first model to meet the Critical cybersecurity threshold in its Preparedness Framework (OpenAI). Critical, in their own definition, means the model can find and...
Standard eviction policies score cached entries on access history alone, ignoring three signals agentic execution hands you for free: recomputation cost, DAG dependency count, and agent invocation frequency (arXiv 2607.20495). Combining them into one scoring function reduces l...
Five tiers collapsed into three AI-native ones (Foundation, Advanced, Prime) effective April 9, with the old Standard/Pro/Pro Plus/Enterprise/Enterprise Plus structure hitting end-of-sale July 1. This is the quiet version of the AI pricing shift. Rather than announce per-outco...
This study investigates dropping speculative decoding's lossless guarantee without any training, quantifying speed-ups against controlled capability drift. Standard spec decoding exactly preserves the sampling distribution. Relaxing it buys latency at a small distributional co...
Using op-schema-aware seeded fuzzing against a high-precision fp64 CPU reference on 24 Triton kernels, 15 correct and 9 intentionally buggy, the method caught all 9 buggy variants and passed all 15 controls across five GPU classes (arXiv:2606.20128). Standard kernel benchmarks...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.